[R] loss of information in pdf plots

2012-04-17 Thread Unger, Kristian, Dr.
Hi there is it possible that pdfs generated using the pdf() function with default settings leads to loss of information? I was plotting copy number changes from Agilent 180k data in form of rectangles (rect()) while each rectangle represents one region of copy number change. When plotting into

Re: [R] loss of information in pdf plots

2012-04-17 Thread Petr PIKAL
Hi Hi there is it possible that pdfs generated using the pdf() function with default settings leads to loss of information? I was plotting copy number changes from Agilent 180k data in form of rectangles (rect()) while each rectangle represents one region of copy number change. When

Re: [R] loss of information in pdf plots

2012-04-17 Thread Unger, Kristian, Dr.
Thanks Petr. Unfortunately this does not help. Kristian Am 17.04.2012 um 14:18 schrieb Petr PIKAL: Hi Hi there is it possible that pdfs generated using the pdf() function with default settings leads to loss of information? I was plotting copy number changes from Agilent 180k data in

Re: [R] loss of information in pdf plots

2012-04-17 Thread Uwe Ligges
On 17.04.2012 14:35, Unger, Kristian, Dr. wrote: Thanks Petr. Unfortunately this does not help. In that case, we need a reproducible example. Uwe Ligges Kristian Am 17.04.2012 um 14:18 schrieb Petr PIKAL: Hi Hi there is it possible that pdfs generated using the pdf() function

Re: [R] loss of information in pdf plots

2012-04-17 Thread Marc Schwartz
On Apr 17, 2012, at 7:08 AM, Unger, Kristian, Dr. wrote: Hi there is it possible that pdfs generated using the pdf() function with default settings leads to loss of information? I was plotting copy number changes from Agilent 180k data in form of rectangles (rect()) while each rectangle

Re: [R] loss of information in pdf plots

2012-04-17 Thread Prof Brian Ripley
There must be loss of accuracy: coordinates in PDF are recorded to finite accuracy (for pdf(), something like 0.01 as I recall). In addition, the R plot engine assumes finite accuracy, a minimum width for lines And as people are pointing out, so do PDF viewers. Do not expect to zoom

Re: [R] loss of information in pdf plots

2012-04-17 Thread Unger, Kristian, Dr.
Thanks Uwe. Please use the following below to reproduce the problem. Also, find the plots as I get them in my Dropbox: http://dl.dropbox.com/u/42940015/Rplot.pdf (plot as I get it when I use the save as... function from the Mac GUI menu. http://dl.dropbox.com/u/42940015/test.pdf (plot using the

Re: [R] loss of information in pdf plots

2012-04-17 Thread Uwe Ligges
As far as I can see, the recommended way of using pdf() plot dev.off() works precisely, just zoom in far enough so that your screen is able to display the gaps, e.g. by zooming in by 6400% Uwe Ligges On 17.04.2012 15:53, Unger, Kristian, Dr. wrote: Thanks Uwe. Please use the following

Re: [R] loss of information in pdf plots

2012-04-17 Thread Unger, Kristian, Dr.
Thanks Uwe, you are right! The reason why it works fine now might be that the borders switch of the rect() function was set to white rather than NA when I sent my initial request. This obviously led to drawing of white borders that overlaid some of the tiny blue or red rectangles. However, it

Re: [R] loss of information in pdf plots

2012-04-17 Thread Uwe Ligges
On 17.04.2012 16:16, Unger, Kristian, Dr. wrote: Thanks Uwe, you are right! The reason why it works fine now might be that the borders switch of the rect() function was set to white rather than NA when I sent my initial request. This obviously led to drawing of white borders that overlaid